A synthetic dynamic image is output by first transforming each frame of a first dynamic image based on a relative positional relationship between each frame of the first dynamic image to create a first transformed dynamic image. Next, each frame of a second dynamic image is transformed based on a relative positional relationship between at least one frame including the first dynamic image and at least one frame including the second dynamic image, and a relative positional relationship between each frame of the second dynamic image to create a second transformed dynamic image. Finally, the synthesized dynamic image is synthesized by aligning, for each frame of the first transformed dynamic image and the second transformed dynamic image, a position in the first transformed dynamic image with a position in the second transformed dynamic image.